UC San Diego

This study presents a reconfigurable retinomorphic infrared imager that can be electrically switched between two complementary modes: a motion mode that delivers streamlined spiking outputs for efficient motion tracking and a static mode that integrates charge to capture high-fidelity images. Mode selection is achieved at the device level by tuning the variable resistance in a photocapacitor-transistor architecture. This work also establishes measurement guidelines for time-derivative sensors to evaluate rate-dependent detection limits. To demonstrate scalability and integration capability, 8 × 8 pixel arrays are fabricated and integrated with silicon integrated circuits, enabling spatiotemporal imaging in dual modes. This device-level approach reduces circuit overhead traditionally required for multimodal operation and offers a compact configuration for lowering energy consumption and facilitating high-density integration in next-generation vision systems.